Over the past few years, the use of chatbots powered by artificial intelligence (AI) has skyrocketed. These conversational agents have become ubiquitous on websites, social media platforms, and messaging apps, helping businesses automate customer support, streamline sales, and improve user engagement.
However, as with any new technology that captures the imagination of investors and entrepreneurs, there are concerns that AI chatbots are experiencing a “tulip mania” moment. This term refers to the Dutch tulip market bubble of the 17th century, where the price of tulips became so inflated that they were being traded for more than a house.
OpenAI recently unveiled ChatGPT-3, an LLM with an astonishing speaking aptitude and can reply to various inquiries. The launch on November 30 this year has created huge anticipation as the AI technology demonstrates the uncanny conversational ability.
Now conscious that a rival had overtaken them regarding artificial intelligence technology supremacy, Google hurriedly issued an internal notice labeled “Code Red,” – alerting their staff.
Google is renowned for its search engine, which holds 84% of the worldwide search market. Their online search utility is so popular that they have given rise to the common term “googling,” meaning to browse the web.
Using its knowledge of the user, Google’s search engine returns helpful links and targeted advertisements in response to the search request posed by the user – and it is aware of more information than needed.
The links provided, although generally useful, require a significant amount of time to examine and locate the precise information one needs.
GPT (Generative Pre-trained Transformer), a search engine from OpenAI, enables users to get direct answers easily without navigating many links like a scavenger hunt. Microsoft, which has its competing search engine, viewed this attribute of GPT to be groundbreaking.
Google declaring “Code Red” for the proposed investment of OpenAI is due to Microsoft’s decision to invest $10 billion, on top of its earlier $1 billion, hoping that GPT will turn around its struggling Bing search engine and gain an edge over Google in the search engine wars.
Google’s LaMDA language model has not yet been released for public testing due to their policy against unauthorized information disclosure. Though it is in development, details regarding its performance are unknown.
On February 7, Google joined the conversation on GPT’s thrilling emergence with a public presentation showcasing LaMDA’s strength.
Unfortunately, LaMDA bombed its job interview. Answering the question, “What exciting discoveries from the James Webb Space Telescope can I explain to my 9-year old?”, LaMDA was unsuccessful.
The LaMDA gave two accurate responses; however, it incorrectly claimed the Webb telescope had snapped photography of a heavenly body located beyond our solar system, being the first to do so–the European Southern Observatory’s Very Large Telescope was responsible for collecting initial photographic evidence.
Upon news of the market value loss, Google’s stock price quickly fell by 8%. This resulted in a total of $100 billion loss for the company.
When Microsoft showcased its language model GPT to reporters, it proved that this system was not as great as most people thought; LaMDA follows close behind. Though there were wrong answers from Bing at the demonstration, this imperfection reflects a marginal flaw peculiar to all LLMs.
Using unparalleled levels of text, Language Modeling (LLM) works by employing statistical values to form coherent sentences from individual words. These LLMs act as advanced text producers in that regard.
Instead of being genuinely ‘intelligent,’ these models operate as automated calculators spitting out words; however, this does not enable them in any way to verify the truthfulness of their strongly stated responses.
AI chatbots are undoubtedly a game-changing technology that has the potential to revolutionize the way businesses interact with their customers. However, as with any new technology, there are risks associated with its use. While AI chatbots can provide businesses with significant benefits, such as customer support automation, improved user engagement, and streamlined sales, challenges must be addressed, such as the need for natural language processing and the potential for miscommunication with customers.
The analogy of “tulip mania” raises an important point about the potential risks of over-hyping AI chatbots. It’s important to approach AI chatbots with a clear understanding of their capabilities, limitations, and potential risks.
While the hype around chatbots may lead to inflated expectations, businesses that approach them with a clear understanding of their strengths and limitations will be better positioned to make the most of their potential.
Source: salon.com
